Research map / Detection and segmentation / Open-set and few-shot detection
Cycle-Consistent Learning for Joint Layout-to-Image Generation and Object Detection
Xinhao Cai, Qiuxia Lai, Gensheng Pei, Xiangbo Shu, Yazhou Yao, Wenguan Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add GDCC cycle fine-tuning (2 epochs) | GeoDiffusion · COCO 2017 L2I fidelity | FID: 20.16 -> 18.02 | Improved |
| add GDCC cycle fine-tuning | GeoDiffusion · COCO 2017 L2I fidelity | YOLO score mAP: 29.1 -> 31.4 | Improved |
| add GDCC instead of plain fine-tuning (ablation) | GeoDiffusion · COCO 2017 L2I fidelity | FID: 20.13 -> 18.02 | Improved |
| add GDCC fine-tuning | Faster R-CNN · COCO 2017 detection | mAP: 37.3 -> 38.5 | Improved |
| add GDCC fine-tuning | ControlNet · COCO 2017 L2I fidelity | FID: 28.14 -> 26.38 | Improved |
| add GDCC fine-tuning | GeoDiffusion · NuImages L2I fidelity | FID: 9.58 -> 7.97 | Improved |
| add GDCC fine-tuning | Faster R-CNN · NuImages detection | mAP: 36.9 -> 37.9 | Improved |
| add layout translation cycle loss (Lgen) (ablation) | GeoDiffusion · COCO 2017 L2I fidelity | FID: 20.13 -> 18.94 | Improved |
| add image translation cycle loss (Ldet) (ablation) | Faster R-CNN · COCO 2017 detection | mAP: 37.5 -> 38.0 | Improved |
| add priority timestep re-sampling w=6 (ablation) | GDCC · COCO 2017 reward strategy | FID: 19.57 -> 18.02 | Improved |
| raise tthre from 50 to 100 (ablation) | GDCC · COCO 2017 reward strategy | mAP: 38.5 -> 37.9 | Hurt |
| add 150k synthetic layouts and images | GDCC · COCO 2017 data efficiency | FID: 18.02 -> 17.16 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Faster R-CNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- GeoDiffusion: Text-Prompted Geometric Control for Object Detection Data Generation (ICLR 2024 · similarity 0.86)
- ODGEN: Domain-specific Object Detection Data Generation with Diffusion Models (NeurIPS 2024 · similarity 0.85)
- Rethinking The Training And Evaluation of Rich-Context Layout-to-Image Generation (NeurIPS 2024 · similarity 0.84)
- Generative Region-Language Pretraining for Open-Ended Object Detection (CVPR 2024 · similarity 0.84)
- Point2RBox-v2: Rethinking Point-supervised Oriented Object Detection with Spatial Layout Among Instances (CVPR 2025 · similarity 0.84)
- HiCo: Hierarchical Controllable Diffusion Model for Layout-to-image Generation (NeurIPS 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
